MCP-Enabled LLM for Meta-optics Inverse Design: Leveraging Differentiable Solver without LLM Expertise

TL;DR

This paper introduces a Model Context Protocol (MCP) framework that enables researchers to perform metasurface inverse design using large language models (LLMs) with minimal programming expertise, by providing dynamic access to verified code templates and documentation.

Contribution

The proposed MCP-assisted framework allows LLMs to autonomously generate inverse design codes by accessing verified resources, improving usability and efficiency in scientific computing tasks.

Findings

Structured prompting outperforms natural language prompting in design quality.

MCP enables access to complex computational tools with only 5 APIs.

Framework reduces the need for programming expertise in inverse design.

Abstract

Automatic differentiation (AD) enables powerful metasurface inverse design but requires extensive theoretical and programming expertise. We present a Model Context Protocol (MCP) assisted framework that allows researchers to conduct inverse design with differentiable solvers through large language models (LLMs). Since LLMs inherently lack knowledge of specialized solvers, our proposed solution provides dynamic access to verified code templates and comprehensive documentation through dedicated servers. The LLM autonomously accesses these resources to generate complete inverse design codes without prescribed coordination rules. Evaluation on the Huygens meta-atom design task with the differentiable TorchRDIT solver shows that while both natural language and structured prompting strategies achieve high success rates, structured prompting significantly outperforms in design quality, workflow efficiency, computational cost, and error reduction. The minimalist server design, using only 5 APIs, demonstrates how MCP makes sophisticated computational tools accessible to researchers without programming expertise, offering a generalizable integration solution for other scientific tasks.